If 2024 is, as I’ve mentioned on occasion, the year of artificial intelligence in the client, Procyon AI Image Generation Benchmark is undoubtedly the tool that many professionals have been waiting for, and as you can already understand from its name, we are talking about a tool designed specifically for evaluating system performance when performing artificial intelligence imaging tasks.

This new tool, which expands UL’s already extensive and highly prestigious portfolio, it relies on Stabble Diffusionmore specifically in versions 1.5 (faster) and XL (capable of creating larger images), specializing in creating images of 512×512 and 1024×1024 pixels, and those responsible confirm in a statement that after its release they report that their plans include adding additional tests in the future to support other performance categories.

In terms of supported specialized technology platforms, since its launch last Monday, the Procyon AI Image Generation Benchmark is now compatible with the three most relevant on the market in terms of client computing. They are the following:

• Intel OpenVINO
• NVIDIA TensorRT
• ONNX Runtime with DirectML

And further, as with regard to the models used, UL confirms that its plans for the future also include add support for more inference engines in the futurethereby expanding its reach.

As I said at the beginning, Procyon AI Image Generation Benchmark will reach out to professionals at the most convenient time, since in 2024 we will see quite a few devices arrive on the market with chips that already integrate specific electronics for computing artificial intelligence, as we have already seen with Intel Core Ultra, without having to resort to online services. But of course there are multiple factors that affect the final performance when performing this type of task.

So, It is very important to have a frame of reference which makes it possible to compare the performance of the mentioned systems in performing a task, as specific and at the same time required by users, as the generation of images with the help of artificial intelligence, a type of activity aimed at replication, in a short period of time. , the current arrival of solutions such as NVIDIA’s Chat With RTX and LM Studio, which we recently informed you about, dedicated to running AI-based chatbots in the client.

Procyon AI image generation benchmark integrates into the flexible UL Procyon Benchmarking Suite modela set of UL applications designed to perform several specialized performance tests, such as the ones we performed in the ASUS Vivobook Pro 15 OLED analysis, which already includes other specific tests to measure the inferential capacity of the tested systems.